摘要
针对当前移动推荐中存在的"信息过载"和推荐质量不高等问题,提出一种基于地理位置和协同过滤相结合的移动推荐算法。算法根据用户与物品间距离对被推荐物品进行预过滤,以缩小推荐范围,并根据用户间的相似度,对预过滤的物品进行偏好预测推荐。实验证明,该算法明显优于基于用户协同过滤推荐算法(UserCF)和基于位置的最近距离均值推荐方法(DARS)。
With the rapid development of O2O E-business, mobile recommendation algorithm became a research focus. Aiming at problems such as "information overload" and poor recommendation quality existing in the current mobile recommendation, the paper put forward a mobile recommendation algorithm based on the combination of geographic position and collaborative filtering. In this algorithm, recommended objects were pre-filtered according to the distance between a user and the object so as to narrow the recommendation scope, and then preference prediction were recommended to the pre-fihered objects according to similarity among users. Experimental results showed that the results obtained through this algorithm obviously prevailed over results from the userbased collaborative filtering recommendation algorithm (UserCF) and the position-based minimum distance average recommendation method (DARS).
出处
《新乡学院学报》
2016年第9期30-33,共4页
Journal of Xinxiang University
基金
安徽省高等学校省级质量工程--大学生创客实验室建设计划项目(2015ckjh139)
安徽省高校优秀青年人才支持计划重点项目(gxyq ZD2016489)
安徽商贸职业技术学院科研项目(2016KYZ04)
关键词
地理位置
协同过滤
移动推荐
O2O
geographic location
collaborative filtering
mobile recommendation
O2O